Mercedes Schneider has a Ph.D. In statistics. She teaches in St. Tammany Parish in Louisiana. Recently, she has discovered that the Louisiana Department of Education has engaged in fancy statistical manipulation of test scores and school grades. The bottom line is that the public is getting spin and hype from that agency, not trustworthy data about student achievement.
In this post, Mercedes explains her efforts to get honest data from the Department. She was thwarted repeatedly.
Louisiana is a beautiful state with wonderful culture and warm people.
For generations, Louisiana has had a reputation as one of the most corrupt state governments in the nation. Plus Ca change, plus c’est la meme chose.
Diane Ravitch's blog 
Edit needed? “The bottom line is that the public is not getting spin and hype from that agency, not trustworthy data about student achievement.” Strike the first “not”?
Strike the word “not.”
The more I hear, the more it seems that the insanity in Louisiana is not an accident or just the result of the abject cluelessness of their politicians. This is the test case for the destruction of not just public education, but education itself. As we have seen in other places in our society, it is the tax dollars of the victims that are being used to destroy those victims. The gun lobby, the Ag lobby and Monsanto, all of these and more are making the 9/11 hijackers look like dilettantes in the way they continue to use our own resources against us, dividing us from what was once at least partially our government.
Being in denial of fake test scores is a part of how you maintain your reputation of being majorly corrupt. It looks like they are proud of that. This is a national cancer that must be irradiated out, cut out or both.
This is not confined to Louisiana. As a simple exercise that does not require a Ph.D. in statistics, simply have a look at the variations in the formulas that states use from year to year to convert their raw scores into final scores. Then make separate plots of the raw scores and the final scores over time. What you’ll find is that the latter show a steady increase, while the former bounce all over the place. Then use a simple program like Excel to do regressions. The final scores typically give you a line heading fairly sharply upward–adequate yearly progress. The raw scores give one that stays flat, declines, or shows very slight increases. Such an exercise makes one wonder what’s going on at the federal level. Does the DOE simply take these reported scores at face value? Does not one apply to states’ reported results even a rudimentary knowledge of statistics, of the level that one might learn in middle school? I am not a statistician by training, but I did this for one state a couple years ago and was shocked at the results. BTW, it’s often said that one can use statistics to say whatever one wants. This is not true. One can use faked statistics to say whatever one wants. There’s a big difference.
I’ve long thought that the project of doing such an analysis of states’ reported scores on their high-stakes tests would make a great dissertation project for some enterprising graduate student of education and that such a project would be quite incendiary.
Here’s another thing. I’ve worked for years in the textbook industry. Back during the Bush, Jr., administration, the DOE instituted requirements that K-12 educational materials be shown, via research, to be effective. So, every K-12 textbook publisher started publishing white papers on the research that proved that their programs were effective. Guess what? Every program was extraordinarily effective, and the stats from the publishers’ research “proved” that this was so. It’s an easy matter to go online and look at these whitepapers, at the publishers’ “research.” A statistician, reading these reports, would have a field day, for the reports are routinely, typically FULL of howlers, misrepresentations, and outright lies.
And, of course, the publishers all have creative writing departments for generating, on demand, the ” research” that “proves” their programs’ effectiveness.
Oh, and BTW, you will often find that the conversion factor used to change raw scores to final scores does not simply change from year to year but from grade level to grade level, in what seems to be a weirdly random manner, at first, within the same subject area. When you plot the two types of scores, you see why. Scores for 4th graders dropped, so you change the conversion factor for that grade level to mask the fact. Abacadabra, annual yearly progress.
Well, considering that educational standards and standardized testing is so fraught with error (See Noel Wilson: “Educational Standards and the Problem of Error” found at:
http://epaa.asu.edu/ojs/article/view/577/700
“A Little Less than Valid: An Essay Review” found at:
http://www.edrev.info/essays/v10n5index.html
http://www.edrev.info/essays/v10n5.pdf )
you can see why the publishers have to obfuscate, prevaricate and outright lie about their “products”.
Thanks, Robert for bringing out these “problems” with the publishing industry!!
Eber, vete!
Just curious–what does that mean?
To what is the “that” referring?
So long as all students have a right to an education, there must be a public school system. Private and corporate education entities simply will not accept all students.